TOXOPLASMOSIS IN THE MARSHALL ISLANDS 317 TABLE 2 Prevalence in seropositive Marshallese of chorioretinal scars compatible with toxoplasmosis By radiation exposure (minimum age, 28 years): PER CENT Rongelap Figure 2, 30-39 40-49 50-59 AGE IN YEARS »60 Percent of negative tests for antibody to Toxoplasma gondii among Marshallese radiation ex- posure categories: open bars = high exposure group (Rongelap, n = 62); black bars = low exposure group (Utirik, n = 98); hatched bars = unexposed ‘“‘comparison” population (n = 100). negative. A “negative” titer was defined as <4 (i.e., negative at serum dilutions < 1:16). The variables analyzed in relation to the Toxoplasma titers included sex, age, island of residence at the time serum was drawn, and history of radiation exposure. Ages were normalized to those listed for 1982. Unless specified otherwise, Rongelaprefers to the highly exposed group, Utirik to the low exposure group, and “comparison” to the matched control group. ° RESULTS Using the x? test of independence between two or more samples, no significant difference in the prevalence of positive serologic tests for toxoplasmosis was detected among the 517 Marshallese tested when analyzed according to sex (df = 1, x? = 2.78, P = 0.99) or the age groups listed in Table 1 (df = 6, x? = 6.98, P = 0.32). There was a significant variation in distribution of positive tests amongthe four islands (df = 3, x? = 8.50, P = 0.04) with Rongelap island having the lowest prevalence of antibody (Table 1). Analysis of the three radiation exposure groups revealed that the Rongelap group hada significantly lower prevalenceof positive tests than did UVtinik (df = 2, x* = 5.95, P = 0.02) or the comparison group (df = 2, x? = 4.69, P = 0.03). The negative tests in the Rongelap group (17.7%) were not clustered around any particular age (Fig. 2). To determineif local environmental factors may (3.4%) - 3/95 5/90 (3.2%) (5.6%) 4/97 (4.1%) Total 14/340 (4.1%) Comparison Other unexposed persons <30 2/58 Utirik By age (all persons tested): 10-19 20-29 30-39 40-49 50-59 > 59 Total 3/65 4/135 4/104 5/60 3/61 0/60 19/485 (4.6%) (3.0%) (3.8%) (8.3%) (4.9%) (0.0%) (3.9%) have decreased the likelihood of acquiring toxoplasmosis while living on Rongelap island, the prevalence of seropositivity by island was recalculated after excluding the Rongelap exposed group. Nosignificant difference was now detected amongthe islands (df = 3, x* = 1.18, P = 0.76) (Table 1). Furthermore, negative titers in unexposed current residents of Rongelap (4 of 68 persons) weresignificantly less common than amongthe exposed (df = 1, x? = 3.83, P = 0.05). In 152 persons from Rongelap, Utirik, and Ma- juro who were born subsequent to the fallout (Ebeye was not tested), negative tests were most common on Majuro, although this was not sta- tistically significant (df = 2, x7 = 2.51, P = 0.29). The meanlog titers of Toxoplasma antibody among seropositive Marshallese, as determined by analysis ofvariance, did not vary significantly with age (df = 6, F = 1.47, P = 0.19) or history of radiation exposure (df = 2, F = 1.92, P=0.15) (Table 1). In addition, mean logtiters ofantibody in the three radiation exposure groups were not significantly different when analyzed by age group at the time of exposure (all P values >0.05). Twenty-eight persons were foundto haveret- inal lesions compatible with toxoplasmosis, of whom 22 had serum collected for Toxoplasma testing. Nineteen of the latter exhibited Toxo- plasma antibody. The prevalence of chorioreti- nal scars which were judged as probably or possibly due to toxoplasmosis was similar among seropositive persons in the three radiation exposure groups and a fourth group composed of unexposed persons not included in the comparison population but over 27 years of age (df =